PETROPHYSICS, VOL. 44, NO. 6 (NOVEMBER-DECEMBER 2003); P. ??–??

; 9 FIGURES

Effect of Capillary, Viscous and Gravity

Forces on Gas-Condensate Mobility

O. Vizika1 and F. Kalaydjian1

ABSTRACT
The dependence of gas and condensate relative perme- number (ratio of viscous to capillary forces) and the Bond
abilities and of critical condensate saturation on the fluid number (ratio of gravity to capillary forces). The spread-
properties (interfacial tensions, densities and wetting ing characteristics of the condensate on the substrate
characteristics), on rock structure and operational parame- (solid surface or water film) are also taken into account.
ters (velocity) is still poorly captured by the reservoir The model is tested against experimental results reported
simulators. This paper proposes a model based on the in the literature. Good agreement is obtained, indicating
dependence of relative permeability, Kr, and condensate that the model correctly captures most of the controlling
mobility on two dimensionless numbers: the capillary parameters.

INTRODUCTION tem will get when the pressure decreases and the phase
envelope is reached, the liquid accumulation and production
Describing flow processes that occur both far from and
are ruled by the balance between three main mechanisms:
close to the wellbore region is a major issue in accurately
gravity segregation, capillary holdup and viscous drag. In
predicting gas-condensate reservoir performance. When a
order to predict gas deliverability, two major features have
gas-condensate reservoir is produced, the pressure draw-
to be studied: the dependence of Kr on the gas-condensate
down leads to the buildup of a liquid bank that becomes pro-
interfacial tension and on the flow conditions.
gressively mobile. Once mobile, this oil bank flows towards For a long time only the effect of interfacial tension on Kr
the producing wells potentially causing impairment. The has been studied, and Kr changes have been attributed to
liquid accumulation that occurs in the vicinity of the pro- rapid interfacial tension changes near to critical point (Asar
duction wells tends to lower the deliverability of the gas due and Handy, 1988; Bardon and Longeron, 1980; and Haniff
to multiphase flow effects. In addition, the gas that is to be and Ali, 1990). More recently, the effect of flow velocity on
produced tends to become lighter due to condensation and Kr has been acknowledged (Henderson et al., 1993 and
therefore is less marketable. Henderson et al., 1995). The investigations have been ori-
Predicting gas-condensate reservoir performance ented toward a dependence of Kr on the capillary number, a
requires an accurate modeling of the flow behavior coupled dimensionless number that includes both the interfacial ten-
with a correct thermodynamic modeling of the various sion and the velocity (Blom et al., 1997; Henderson et al.,
processes involved. Once the liquid segregates, the way in 1995; Kalaydjian et al., 1996; Schechter and Haynes, 1992).
which the densities of the two phases start diverging and the Another parameter that has been extensively studied is
gas/liquid interfacial tension builds up depends on the ther- the critical condensate saturation, Scc. Determination of this
modynamic properties of the gas-condensate system. value has been controversial. Values ranging between zero
Hence, depending on how close to the critical point the sys- and 50% pore volume have been reported (Barnum et al.,

Manuscript received by the Editor March 10, 2003; revised manuscript received July 28, 2003.
1
Institut Français du Pétrole, Rueil-Malmaison, France.
©2003 Society of Petrophysicists and Well Log Analysts. All rights reserved.

1 PETROPHYSICS November-December 2003

 Effect of Capillary, Viscous and Gravity Forces on Gas-Condensate Mobility

1995 and Morel et al., 1992). This saturation is the mini- assumed to represent the effect of capillary forces, while
mum liquid saturation above which the condensate viscous forces intervene in the Darcy equations and are rep-
becomes mobile, corresponding to a non-zero condensate resented by the Kr. Phase distribution and displacement
relative permeability. These low values of the condensate Kr within the porous medium depend on a complex combina-
control gas-condensate segregation and impact the phase tion of those two forces. Kr was found to depend on interfa-
distribution and the condensate ring buildup. To correctly cial tension (Asar and Handy, 1988; Bardon and Longeron,
predict gravity segregation for near-critical systems, it is 1980 and Haniff and Ali, 1990), a typical capillary parame-
necessary to compare viscous, capillary and gravity forces. ter, and on the flow velocity. When the interfacial tension
To this end two relevant dimensionless numbers are used:
vanishes, the capillary resistance to flow decreases. There-
the capillary number (Ca, ratio of the viscous to the capil-
fore, the curvature of the Kr curves, which expresses for a
lary forces and predominant close to the well) and the Bond
given fluid the capillary effect induced by the presence of
number (Bo, ratio of the gravity to the capillary forces and
predominant far from the well). More specifically, three another fluid, decreases as well. Kr curves become straight
configurations can be identified: 1) near-wellbore region: lines either when miscibility is approached or when flow
high velocity, high interfacial tension (high Ca, low Bo); 2) rate is increased enough to make capillary forces negligible.
reservoir: low velocity, intermediate interfacial tension Therefore, it is anticipated that Kr curves depend on the cap-
(low Ca, high Bo); and 3) near-critical reservoir: low veloc- illary number, Ca, a dimensionless number defined as the
ity, low interfacial tension (high Ca, high Bo). ratio of the viscous to the capillary forces:
The behavior of liquid condensate on the water/gas or
mn
solid/gas interface is a key factor. Complete wetting of the Ca = . (1)
condensate on solid or water (spreading) would favor s
hydraulic continuity leading to low liquid saturations.
Dullien (1992) proposed another capillary number that
There is evidence that for near-critical systems, the conden-
includes the characteristics of the porous medium in addi-
sate phase perfectly wets either the rock or the water phase
tion to those of the fluids
covering the pore surface (Ragil et al., 1998). However, as
the pressure decreases towards the well, a wetting transition Ca. l
may occur that would render the condensate phase only par- CA = 4 , (2)
R
tially wetting on water. This would favor its trapping by
capillary forces. where l is the core length and R is a typical size of a pore
The objective of this paper is to present a model for gas- radius. It is chosen to be equal to the pore size corresponding
condensate Kr as a function of the capillary number and to to the entry pressure deduced from the mercury intrusion
demonstrate that apparently contradictory laboratory curve; that is R = 2.s.cos(q)/P*, with P* being the pressure
results on the dependence on interfacial tension and flow at which the mercury starts invading the porous medium.
rate, separately considered, can be reconciled. This is When this capillary number is equal to unity, viscous forces
achieved, within the framework of the Darcy description of balance capillary forces. When CA ³ 1, viscous forces are
multiphase flow in porous media (relatively far from the dominant.
well so that inertial effects are not an issue), by introducing It is worth checking this assumption by comparison with
a unique dependence of Kr on the capillary number. Also, Bardon and Longeron’s data (1980). These authors report a
modeling of the combined effect of gravity and capillary significant change in Kr for an interfacial tension smaller
forces on the critical condensate saturation is introduced. than 0.04 mN/m. The tests were carried out in a 40-cm long
The model takes into account the spreading characteristics Fontainebleau sandstone with 82-mD permeability and
of the condensate on the substrate (solid surface or water approximately 10% porosity. The velocity was about 20
film). cm/h. According to the authors, the shape of the Kr curves
did change for a capillary number equal to 0.4 10–4 (stan-
RELATIVE PERMEABILITY dard formulation). To calculate the value of the capillary
AND CAPILLARY NUMBER number CA, it is necessary to know the pore size corre-
As seen in the previous section, the interfacial tension sponding to the entry pore pressure. This pore size can be
alone cannot adequately parameterize the relative perme- calculated using the rough estimate Ö(8k/f), which gives an
abilities. The displacement of several phases in a porous average value of about 2 microns and thus a pore radius cor-
medium is governed by two different forces: viscous and responding to the entry pressure of about 10 mm. This gives
capillary (in the absence of gravity). Capillary pressure is CA = 6.4. This value of CA, greater than unity, is thus con-

November-December 2003 PETROPHYSICS 2

 Vizika and Kalaydjian

sistent with dominant viscous forces that tend to decrease over length scales that may range from the nanometer to a
the curvature of Kr curves. few microns and have the same surface fractal dimension
Another example is that found in Henderson et al. DS. DS takes values between 2 and 3, in which 2 character-
(1995). Gas capillary numbers ranging between 0.18E-5 izes smooth clay-free rocks, while values close to 3 are
and 0.14E-4 were reported. The experiments were per- characteristic of strongly clayey sandstones. This fractal
formed on a core 61 cm long, of 92-mD permeability,
dimension can be easily determined from the capillary pres-
19.8% porosity and an irreducible water saturation of 0.264.
This gives CA > 1 (1.96 < CA < 7.86), and explains why sure curve at low wetting phase saturation (Davis, 1989) or
the authors found an effect of the flow rate, and thus of the using sophisticated measurement methods as the X-ray or
capillary number on the relative permeabilities since all the neutron scattering techniques (Bale and Schmidt, 1984),
tests were performed in the domain of dominant viscous improving the measurement accuracy considerably.
forces. The modeling approach proposed here has been inspired
Another example is the work presented by Asar and by the de Gennes (1985) paper on the partial filling of a
Handy (1988). They performed experiments in a Berea fractal structure by a wetting fluid (Lenormand, 1990). The
sandstone sample one ft long, with 20% porosity and internal surface of a pore is assumed to be a fractal surface;
193-mD permeability. Four interfacial tension values rang- consequently, a perfectly wetting phase remains always
ing from 0.03 to 0.82 mN/m were tested. Even though only continuous. The isotropic fractal surface is modeled as a
the range of the applied pressure drops is given, and calcula-
bundle of parallel capillary tubes with a fractal cross sec-
tion of the exact CA for each experiment is not possible, it is
tion. The cross section of each tube is constructed by an
seen that CA ranges between 0.5 and ~80. This explains
why, for the highest interfacial tension (s = 0.82mN/m), iterative process, by dividing the half perimeter of a circle in
capillary effects are dominant and the measured Kr curves h parts and replacing each part by half a circle (Figure 1). At
approach those obtained for conventional gas-oil flood, each step k of the process, Nk new grooves are created with
while for s = 0.03mN/m, viscous effects are dominant and radius Rk and total cross-sectional area Ak; these characteris-
Kr curves approach straight lines. tics are given as a function of the initial tube radius R0 by the
following relationships :
MODELING OF THE RELATIVE PERMEABILITY
R k = R 0 (p / h ) k (3)
(Kr) AND THE CRITICAL CONDENSATE
SATURATION (SCC) Nk = h k (4)
Porous medium model 1
. Ak = pR 02 ( p 2 / h ) k . (5)
It is well known that the transport properties in a porous 2
medium depend strongly on pore structure geometry. Sev- It can be easily shown that the perimeter of a section is given
eral papers proved that sedimentary rocks are one of the by
most extensive natural fractal systems (Katz and Thomp-
son, 1985 and Radlinski et al., 1999). They demonstrated L / L0 = ( R / R 0 ) (1-D L ) , (6)
that the pore volume and the pore-rock interface is fractal
where L0 is the perimeter of the main tube and DL is the lin-
ear fractal dimension associated with the perimeter (linear
fractal dimension, DL = DS – 1) given by
Wate r ln h
Oil DL = . (7)
ln h / p
Gas R1
R2 R0 Capillary pressure and Relative permeability
At equilibrium, all tubes with size smaller or equal to Rk,
where Rk is given by Laplace’s law: Pc = 2s/Rk are occupied
by the wetting fluid, and larger tubes by the non-wetting
one. Thus, the wetting fluid saturation is given as the sur-
FIG. 1 The fractal pore model and the phase distribution within face of the tubes occupied by the wetting phase to the total
it. cross section,

3 PETROPHYSICS November-December 2003

 Effect of Capillary, Viscous and Gravity Forces on Gas-Condensate Mobility

¥
saturations. If CA increases (CA ~ 1), viscous forces start
åA k
æ Rk ö
( 2- D L ) being important, and for high CA (CA > 1), they become
Sw = k
¥
=ç ÷ (8) dominant.
è R0 ø
åA k
For a given macroscopic capillary number, different flow
0 regimes may exist inside the porous medium at the pore
level. In the bundle of capillaries for example, all tubes are
and the correlation between capillary pressure and wetting- subject to the same pressure gradient. Locally, however, dif-
phase saturation is given by ferent flow velocities develop depending on the tube radius.
2s 1/ (D L -2) These different velocities lead to different local capillary
Pc = Sw (9) numbers, meaning that flow may be capillary-dominated in
R0
the smallest pores, while simultaneously viscous-
To simplify calculation of Kr, the grooves are replaced by dominated in the rest. That means that for a given macro-
capillary tubes of the same diameter and parallel to the scopic or bulk capillary number, CA0, we can define a
direction of flow. To calculate Kr, Poiseuille’s law is applied threshold condensate saturation, Stc, below which flow is
in each capillary of the bundle. If the flow rate in a single capillary-dominated and above which viscous forces pre-
tube of radius Rk is given by dominate. Suppose that in the bundle of capillaries model,
Stc occupies the smallest pores from R¥ to Rk. The capillary
pR k
4

Qk =- ÑP , (10) number in the tubes k is CAk. From the definition of CA

8m (equation (2)) it is seen that
then the Kr of the wetting phase, which occupies the smallest CAk R k
= , (12)
tubes, is calculated as CA0 R 0
¥

åQ k Nk which can be combined with equation (8) to express in a dif-

K rw = k
=S ( DL- 4) / ( DL- 2)
. (11) ferent way the wetting fluid saturation
¥ w

åQ k Nk æ CAk ö
( 2- D L )

0 S c =ç ÷ . (13)
è CA0 ø
We have assumed that the wetting fluid flows down to
zero saturation and that there is no irreducible or immobile It is now obvious that the threshold condensate satura-
wetting-phase saturation. However, irreducible phase satu- tion is the one for which CAk = 1.
ration can be easily introduced in the calculations as pre-
sented elsewhere (Vizika, 1993). This model, when applied S tc = (CA0 ) (D L -2) . (14)
to film flow and low wetting-phase saturations, allows esti- For a given pore structure (given DL), this expression
mation of Kr curves for a range of saturation values where
permits calculation of the part of the wetting fluid flowing
reliable experimental results are hard to obtain (Kalaydjian
by capillary-dominated flow as a function of the macro-
et al., 1997 and Moulu et al., 1997).
scopic capillary number. Figure 2 shows the dependence of
Modeling of gas-condensate relative Stc on CA for different fractal dimensions. It is seen that the
permeability as a function of CA higher the capillary number, the lower the threshold con-
densate saturation. It is also seen that, for a given CA, Stc
Threshold condensate saturation increases with DL. This means that, for the same macro-
Here we use the model described above to calculate gas- scopic flow conditions, the more fractal the pore structure,
condensate Kr and the impact of capillary number on it. the higher the wetting-phase saturation (condensate) sub-
There exists experimental evidence that near the critical ject to a capillary flow regime. In other words, in a clayey
point the condensate is the wetting phase that spreads spon- sandstone (high DL), the wetting phase maintains a low
taneously on the solid surface (Williams and Dawe, 1989). mobility up to rather high saturations, even at high capillary
During a depletion, and as the pressure goes down, the con- numbers where viscous forces would be expected to pre-
densate saturation builds up. As the wetting phase pro- dominate.
gresses, the condensate occupies first the surface roughness
and the smallest pores. If CA is low (0 < CA < 1), capillary Condensate relative permeability
forces will be dominant for the whole range of condensate The wetting fluid Kr is given by equation (11) up to Stc.

November-December 2003 PETROPHYSICS 4

 Vizika and Kalaydjian

The fluid above the threshold saturation, Stc, flows in the biggest pores with a Krg proportional to the flowing gas
larger tubes by viscous-dominated flow. The respective saturation,
relative permeability for Sc > Stc is taken proportional to the
flowing saturation (straight lines). In calculating wetting Sg
Krg = Krg ( S tc ) . (16)
fluid relative permeability with the bundle of tubes model, 1- S tc
the assumption is made that flow occurs independently in
each tube. This is particularly relevant at rather low liquid It is seen that, in addition to Krc, Krg depends on the
saturations. porous medium (through DL) and on the capillary number
A sample calculation of the condensate (wetting phase) (through the dependence on Stc).
Kr and its evolution with CA is given in Figure 3. A DL = 1.4 A sample calculation of the gas (non-wetting phase) rela-
has been used in these calculations. This value is represen- tive permeability and its evolution with CA is given in Fig-
tative of a weakly clayey sandstone. The Krc curves for three ure 4. As for Krc (Figure 3), a linear fractal dimension of 1.4
different capillary numbers (CA = 3, 10 and 100) are pre- has been used in these calculations. The Krg curves for three
sented in this figure. For comparison purposes, the curve for different capillary numbers (CA = 3, 10 and 100) are plot-
purely capillary-dominated displacement is also given. It is ted. Also the curve for purely capillary-dominated displace-
seen that as viscous forces become important (increasing ment is given for comparison. As viscous forces increase
CA), the part of the fluid flowing under capillary-dominated (increasing CA), the Stc decreases and the saturation range
flow is reduced, and Kr increases considerably. where gas permeability is proportional to Sg increases.

Gas relative permeability Comparison with experiments

Gas as the non-wetting phase occupies the bulk of the The proposed model is now tested against experimental
pores. To calculate its relative permeability, it is assumed measurements of near-critical fluid Kr as reported in the lit-
that, if the condensate occupies all tubes with radius smaller erature. Kr curves at near-critical conditions have been
than Rk, gas flows as though it was in a single pore with reported by Schechter and Haynes (1992). The experiments
radius Rg = R0 + R1 + … + Rk (Moulu et al., 1997). Thus for (injections at an average flow rate of 0.098 cm3/s, for three
low condensate saturations (Sc < Stc), Krg is given by interfacial tensions: 0.1, 0.006 and 0.002 mN/m) were per-
formed in a Clashach sandstone of 30 cm in length, with a
Krg = Krg max (1- S c1/ (2-D L ) ) .
4
(15) permeability of 200 mD and a porosity of 18%. These data
allow calculation of the CA numbers from the Ca values
Krgmax is the maximum value of the gas Kr in presence of given by the authors: for Ca 3 ´ 10–4, 4 ´ 10–3 and 10–2 we
the other phase. For lower Sg, (Sc > Stc), gas circulates in the

FIG. 2 Threshold condensate saturation as a function of the FIG. 3 Effect of capillary number on the condensate relative
macroscopic capillary number for various fractal dimensions. permeability (DL = 1.4).

5 PETROPHYSICS November-December 2003

 Effect of Capillary, Viscous and Gravity Forces on Gas-Condensate Mobility

get the respective CA values 12, 133 and 400. Then Stc, Krc factory agreement is observed considering the uncertainties
and Krg have been calculated with the model by taking DL in the experimental Kr data that have been obtained by
equal to 1.65. This value, representative of a clayey sand- numerical fitting of displacement and production data.
stone (Katz and Thompson, 1985), has been measured in a
Clashach sample of comparable permeability to that used in Gravity segregation and critical condensate saturation, Scc
Schechter and Haynes’ experiments. A comparison between The same model for the porous medium can be used to
Schechter’s measurements and the present model’s predic- calculate the critical condensate saturation for mobility as a
tions are given in Figures 5 and 6. For this example, a good function of the spreading and interfacial properties of the
agreement between experiment and calculations is fluid system, and the characteristics of the pore structure.
obtained. The model successfully predicts gas-condensate We need first to introduce the Bond number, which is
Kr as a function of the macroscopic capillary number. It also defined as the ratio of the gravity to capillary forces, in a
successfully predicts the threshold saturation Stc, below way equivalent to that proposed by Schechter et al. (1994)
which the reduction of condensate mobility becomes sig- (e.g., defined at the macroscopic scale).
nificant. It is worth noting that there is no adjustable
parameter in the model. The only parameter that has been
assigned an arbitrary value (not reported in Schechter and
Haynes, 1992) is the fractal dimension DL.
Bardon and Longeron (1980) have also reported Kr
measurements at different and very low interfacial tensions.
For flow rates 0.011, 0.01 and 0.012 cm3/s and interfacial
tensions 0.065, 0.038 and 0.0014 mN/m respectively, the
Ca numbers calculated by the authors were: 0.19 ´ 10–4,
0.38 ´ 10–4 and 0.11 ´ 10–2. The CA numbers calculated as
indicated above are 2.5, 6 and 168. Comparison between the
experiments and the calculations is given in Figures 7 and 8
for the condensate and the gas Kr, respectively. The fractal
dimension has been taken equal to 1.3. This value is repre-
sentative of a very weakly clayey sandstone as was the one
used in the experiments (Fontainebleau sandstone). A satis-

FIG. 5 Krc: Comparison between the model and Schechter and

Haynes’ experimental results.

FIG. 4 Effect of capillary number on the gas relative permeabil- FIG. 6 Krg: Comparison between the model and Schechter and
ity (DL = 1.4). Haynes’ experimental results.

November-December 2003 PETROPHYSICS 6

 Vizika and Kalaydjian

DrgRl ous phase trapping in the smallest tubes where capillary

Bo = (17)
s forces are very strong, and b) discontinuous phase trapping,
when S < 0 and the condensate is in disconnected form.
It is considered here that Scc is the minimum saturation at
which the condensate is continuous and that the conditions Spreading condensate trapping (S = 0)
are such that gravity forces, favoring segregation, are domi-
In absence of water or if the condensate spreads on the
nant over the capillary forces that favor condensate trap-
water, (S = 0), films are spontaneously formed at the
ping. That means that the condensate has first to form con-
moment of condensate apparition. These films automati-
tinuous films on the solid substrate or on the water, if inter-
cally ensure hydraulic continuity. The continuous phase
stitial water is present in the porous medium. These con-
trapping is related to the Bond number. In the bundle of cap-
tinuous films may still be immobile if gravity forces are not
illaries model, introduced above, the condensate occupies
strong enough compared to capillary forces. Scc is the satu-
the smallest pores, from R¥ to Rk in absence of water, or
ration above which the condensate Kr takes a non-zero
from Rwi to Rk, if the tubes from R¥ to Rwi are occupied by
value. It is to be distinguished from the threshold conden-
irreducible water. The Bond number in the tubes k is Bok.
sate saturation, Stc, that determines the saturation below
From its definition, it results in
which the condensate has a reduced though finite mobility.
In a water-wet porous medium, in presence of irreducible Bok R k
water saturation, a relevant parameter is the spreading coef- = . (18)
Bo0 R 0
ficient of oil (condensate) on water: S = swg – (swc + scg). At
equilibrium, S is always negative or equal to zero. It has If Bok > 1, gravity forces predominate and Rk tubes are emp-
been already verified through numerous studies on three- tied under gravity segregation. The critical condensate satu-
phase displacement that fluid distributions depend on the ration for mobility of a continuous fluid, Scc/c, is the one for
sign of the spreading coefficient of oil on water in presence which Bok = 1 and is given by
of gas. For S = 0, oil (condensate) forms a film spontane-
ously on water in presence of gas. Spreading oil films assure S cc/ c = ( Bo0 ) (D L -2) - S wi . (19)
hydraulic continuity of the oleic phase and can lead to high
recoveries. For S < 0, oil does not spread on water in pres- Non-spreading condensate trapping (S)
ence of gas, and a finite contact angle, q, is formed between
If S < 0, in condensate apparition small droplets are
the gas-condensate and water-condensate interfaces. The
formed. If the substrate is flat, these droplets form an angle
above hold rigorously only for a flat solid surface. Wetting
q for which
on a rough surface is very different, and apparent contact
angles depend both on the wetting properties of the flat sur- S
face and on the structure of the roughness (Onda et al., cos q =1 + . (20)
s cg
1996).Two types of trapping can be considered: a) continu-

FIG. 7 Krc: Comparison between the model and Bardon and FIG. 8 Krg: Comparison between the model and Bardon and
Longeron’s experimental results. Longeron’s experimental results.

7 PETROPHYSICS November-December 2003

 Effect of Capillary, Viscous and Gravity Forces on Gas-Condensate Mobility

On a fractal surface, they form an apparent contact angle, A model has been proposed to calculate gas-condensate
qapp, related to q with the following expression (Onda et al., Kr curves as functions of the capillary number. The model
1996) includes the structural characteristics of the porous medium
D L -1
through its fractal dimension. It predicts the modification of
æ R0 ö the Kr curves as the capillary number changes (velocity or
cos q app =ç ÷ cos q , (21)
è Rk ø interfacial tension changes). The model has been tested
against experimental results reported in the literature and a
where R0 and Rk are the upper and the lower limits of fractal satisfying agreement has been obtained.
behavior, respectively. If cosqapp = 1, the non-spreading con- A threshold condensate saturation, Stc, can be predicted
densate (cosq < 1) remains continuous. The fractal surface below which the condensate mobility is significantly
induces wetting of a non-spreading liquid. Thus the criterion reduced, even though finite. Stc may be very high in highly
for spreading phase mobility has to be applied, as explained fractal (very clayey) sandstones. It decreases with increas-
above. The critical condensate saturation is given by equa- ing capillary number.
tion 19. The critical saturation for condensate mobility, Scc,
When cosqapp < 1, the condensate forms lenses and increases with increasing interfacial tension (decreasing
remains disconnected. It is considered that the disconnected Bond number) and fractal dimension. Non-spreading con-
phase remains immobile and is not subject to gravity segre- densates would be subject to severe trapping, increasing
gation. The critical saturation for discontinuous liquid with increasing fractal dimension and decreasing spreading
mobility, Scc/d, is the one occupying tubes up to Rk for which coefficient.
cosqapp becomes equal to 1. It is given by the following
expression as a function of the spreading and interfacial NOMENCLATURE
properties of the system and the characteristics of the pore
structure here expressed by the fractal dimension DL: Ak Area of the grooves of step k
Ca Standard capillary number (= mv/s)
1
S cc/ d = 2- D L
(1- (cos q) 2-D ) . L
(22) CA Capillary number (= 4 Cal/R)
Bo Bond number (= DrgRl/s)
p D L -1
DL Linear fractal dimension
Then the critical saturation for mobility of a non- DS Surface fractal dimension
spreading condensate is the sum of the trapped continuous g Acceleration of gravity
(if any) and trapped discontinuous liquid. Figure 9 shows k Permeability
the critical condensate saturation for a discontinuous phase
(S < 0, cosqapp < 1) for various values of the fractal dimen-
sion. Scc/d can take extremely high values for strongly non-
spreading fluids (q =90°, cosq = 0) and highly fractal pore
space (high DL).

CONCLUSIONS
Flow rate and interfacial tension dependence of relative
permeability are explained on the basis of the competition
between capillary and viscous forces. Kr curves are shown
to deform and approach straight lines when viscous forces
overcome capillary forces. This can be achieved either by
decreasing the interfacial tension or increasing the flow
rate. To account for these effects, a capillary number, CA,
introduced by Dullien (1992) is used. It includes, in addi-
tion to fluid properties, porous medium properties. CA
allows precise definition of the flow rate threshold above
which (or interfacial tension below which) Kr curves start to
deform. This hypothesis has been validated on experimen- FIG. 9 Krg: Effect of cosq (negative spreading coefficient) on
tal results obtained either by modifying the interfacial ten- the critical condensate saturation of a discontinuous phase for
sion or the flow rate. various values of the fractal dimension.

November-December 2003 PETROPHYSICS 8

 Vizika and Kalaydjian

Kr Relative permeability recovery reduction due to condensation, SPE 30767, in SPE

Kri Relative permeability of fluid i Annual Technical Conference and Exhibition, Dallas, Texas.
Krgmax maximum gas relative permeability Blom, S. M.P ., Hagoort, J., and Soetekouw, D. P. N., 1997, Rela-
l Porous medium length tive permeability at near-critical conditions, SPE 38935: Soci-
ety of Petroleum Engineers, presented at the 1997 SPE Annual
L Perimeter of a section of the fractal object
Technical Conference and Exhibition.
Nk Number of objects of step k Davis, H. T., 1989, On the fractal character of the porosity of natu-
Pc Capillary pressure ral sandstone: Europhysics Letters, vol. 8, no. 7, p. 629–632.
Qk Flow rate in tube of radius Rk de Gennes, P. G., 1985, Partial filling of a fractal structure by a wet-
R Entry pore size ting fluid, in Adler, Fritzche, Ovshinsky, eds., Physics of Disor-
R0 Initial capillary tube radius in the fractal dered Materials, Plenum Publishing Corporation, p. 227–241.
pore model Dullien, F.A.L., 1992, Fluid transport and pore structure, 2nd Edi-
Ri Capillary radius (i = 1,2,...k step of the fractal tion, Academic Press Inc.
construction) Haniff, M. S., and Ali, J. K., 1990, Relative permeability and low
tension fluid flow in gas condensate systems: SPE 20917: pre-
Rg Radius for gas flow in the fractal pore model
sented at the 1990 SPE European Petroleum Conference.
Rwi Maximum tube radius occupied by Henderson, G. D., Danesh, A., Tehrani, D. H., and Peden, J. M.,
irreducible water 1993, An investigation into the processes governing flow and
S Saturation recovery in different flow regimes present in gas-condensate
S Spreading coefficient reservoirs: SPE 26661, presented at the 1993 SPE Annual
Si Saturation of fluid i Technical Conference and Exhibition.
Swi Irreducible water saturation Henderson, G. D., Danesh, A., Tehrani, D. H., Al-Shaidi, S., and
Scc Critical condensate saturation Peden, J. M., 1995, Measurement and correlation of gas con-
Scc/c Critical condensate saturation for densate relative permeability by the steady-state method, SPE
continuous phase 30770, presented at the 1995 SPE Annual Technical Confer-
ence and Exhibition.
Scc/d Critical condensate saturation for
Kalaydjian, F. J-M., Bourbiaux, B. J., and Lombard, J-M., 1996,
discontinuous phase Predicting gas-condensate reservoir performance: how flow
Stc Threshold condensate saturation parameters are altered when approaching production wells:
v Velocity SPE 36715, presented at the 1996 SPE Annual Technical Con-
ÑP Pressure gradient ference and Exhibition.
Dr Density difference Kalaydjian, F. J-M., Moulu, J-C., Vizika, O. and Munkerud, P-K.,
h Number of new tubes created at each step 1997, Three-phase flow in water-wet porous media: gas/oil
q Contact angle between gas-condensate and relative permeabilities for various spreading conditions: Jour-
water-condensate interfaces nal of Petroleum Science Engineering, vol. 17, 275–290.
Katz, A. J. and Thompson, A. H., 1985, Fractal sandstone pores:
m Viscosity
implications for conductivity and pore formation: Physical
s Interfacial tension Review Letters, vol. 54, no. 12, p. 1325–1328.
f Porosity Lenormand, R., 1990, Gravity-assisted inert gas injection: micro-
model experiments and model based on fractal roughness: pre-
Subscripts sented at the 1990 European Oil and Gas Conference.
c, g, w Condensate, gas, wetting phase Morel, D. C., Lomer, J.-F., Morineau, Y. M ., and Putz, A. G., 1992,
Mobility of hydrocarbon liquids in gas condensate reservoirs:
interpretation of depletion laboratory experiments: SPE 24939,
REFERENCES presented at the 1992 SPE Annual Technical Conference and
Asar, H., and Handy, L.L., 1988, Influence of interfacial tension on Exhibition.
gas-oil relative permeability in a gas-condensate system: Moulu, J. C., Vizika, O., Kalaydjian, F., and Duquerroix, J-P.,
SPERE, February, 257 p. 1997, A new model for three-phase relative permeabilities
Bale, H. D., and Schmidt, P. W., 1984, Small-angle X-ray- based on a fractal representation of the porous medium: SPE
scattering investigation of submicroscopic porosity with fractal 38891, presented at the 1997 SPE Annual Technical Confer-
properties: Physical Review Letters, vol. 53, 596 p. ence and Exhibition.
Bardon, C. and Longeron, D., 1980, Influence of very low interfa- Onda, T., Shibuichi, S., Satoh, N., and Tsujii, K., 1996, Super-
cial tensions on relative permeability: Society of Petroleum En- water-repellent fractal surfaces: Langmuir, vol. 12, no. 9, p.
gineers Journal, October, p. 391–401. 2125–2127.
Barnum, R. S., Brinkman, F. P., Richardson, T. W. and Spillette, A. Radlinski A. P., Radlinska, E. Z., Agamalian, M. Wingall, G. D.,
G., 1995, Gas condensate reservoir behaviour: productivity and Lindner, P., and Randl, O.G., 1999, Fractal geometry of rocks:
Physical Review Letters, vol. 82., no. 15, p. 3078–3081.

9 PETROPHYSICS November-December 2003

 Effect of Capillary, Viscous and Gravity Forces on Gas-Condensate Mobility

Ragil, K., Bonn, D., Broseta, D., Indekeu, J., Kalaydjian, F., and ABOUT THE AUTHORS
Meunier, J., 1998, The wetting behavior of alkanes on water:
Journal of Petroleum Science Engineering, vol. 20, p. 177–183.
Schechter, D. S. and Haynes, J. M., 1992, Relative permeabilities
of a near critical binary fluid: Transport in Porous Media, vol.
9, p. 241–260.
Schechter, D. S., Zhou, D., and Orr, Jr., F. M., 1994, Low IFT
drainage and imbibition: Journal of Petroleum Science Engi-
neering, vol. 11, p. 283-300.
Vizika, O., 1993, Effect of the spreading coefficient on the effi-
ciency of oil recovery with gravity drainage: presented at the
1993 Symposium on Enhanced Oil Recovery, 205th National
Meeting of ACS.
Williams, J. K. and Dawe, R. A., 1989, Near-critical condensate
fluid behavior in porous media – a modeling approach: SPE, Olga Vizika has been working with IFP for thirteen years. She
May, p. 221-227. worked on relative permeability modeling for various processes as
depressurization, gas injection, WAG, gas condensate production.
Her current research interests also include modeling of the
transport properties of carbonate rocks and slightly consolidated
porous media. She served as chairman of the Technical Committee
of the 2003 Society of Core Analysts Symposium. She holds a BS
and PhD degree, both in Chemical engineering, from the U. of
Patras, Greece. She is Head of Petrophysics Department in the
Reservoir Engineering Research Division of IFP.

François Kalaydjian has been working with IFP for 17 years.

His research interests include the modelling and simulation of
various recovery processes such as imbibition in fractured
reservoirs, gas injection, WAG and production of gas condensates.
He has authored and co-authored more than 50 articles related to
these research topics. He served on various SPE program
committees and was appointed chairman of the Program
Committee of the SPE Europec 2000 “Integrated Reservoir Man-
agement.” F. Kalaydjian holds an engineering degree from the
School of Mines in France, and a PhD degree in Physics from the
University of Bordeaux (France). After five years as managing
director of the reservoir engineering research division, he is
currently deputy managing director of the Exploration and
Reservoir Engineering Technical Business Unit.

November-December 2003 PETROPHYSICS 10